Correlation and Conversion of the Normalized Constant Score and Patient-reported Outcomes in Shoulder Arthroplasty

There is a paucity of literature describing correlations between the Constant score and commonly used patient-reported outcome measures (PROMs) in shoulder arthroplasty. This study aims to establish the correlation of and conversion between the normalized Constant score (nCS) to commonly used patient-reported outcomes in shoulder arthroplasty patients. METHODS: A retrospective review of a prospectively collected research database was done from 2003 to 2014. Inclusion criteria were primary shoulder arthroplasty (anatomic or reverse) and minimum 2-year follow-up. Preoperative and postoperative outcomes scores (1-year and 2-year) were prospectively collected and included the nCS, Simple Shoulder Test (SST), Shoulder Pain and Disability Index (SPADI), and American Shoulder and Elbow Surgeons (ASES) Score. The nCS was correlated with PROMs (Pearson correlation for SPADI/ASES scores and Spearman correlation for SST). RESULTS: A total of 762 shoulders in 721 patients with 1661 individual clinical encounters were included. The average age of included patients was 67.7 years, 48% of patients being female. Reverse total shoulder arthroplasty (RSA) was more commonly done compared with anatomic total shoulder (aTSA) (57.3% vs. 42.7%, respectively). The nCS correlated strongly with the PROs overall: ASES (0.893, P < 0.001), SPADI (−0.896, P < 0.001), and SST (0.873, P < 0.001). Correlations were similar overall between aTSA and RSA. Preoperative correlations between the nCS and the PROMs were on the high side of moderate correlation with RSA (R = 0.621 to 0.659) and on the low side of a strong correlation with aTSA (R = 0.704 to 0.705) except for the SST (R = 0.608). The 1- and 2-year postoperative time points all showed strong correlation of the nCS with PROMs, except the SST in RSA (R = 0.694). CONCLUSION: The nCS shows high correlation with ASES score, SPADI, and SST in shoulder arthroplasty patients. This suggests that PROMs may be able to be used for shoulder function assessment without the need for physician input for large cohort studies. In addition, the conversion equations generated here may provide utility in the evaluation of shoulder arthroplasty outcomes across studies in systematic reviews and meta-analyses.